"Adam M. Costello" wrote:
> Exactly which "reasonable" characters are you talking about? All UCS character codes are valid for something, and anything can be stored in DNS as an owner domain name (think TXT) or as data (SOA, RP, or as-yet-invented). The problem with the current spec is that case-folding and normalization will impose restrictions on these assignments. This is inappropriate for a generic i18n domain name label syntax. On a secondary note, normalization and lowercasing are problems having to do with hostnames, and those should be enforced by delegation and application rules. Note well: the existing DNS does not enforce hostname rules anywhere else, we should be mimic'ing this behavior. The hostname rules should only define, the codecs should only encode/decode, and the apps/protos/formats should enforce. > The prohibited list in nameprep is pretty small, and they're all > characters that you'd have to be somewhat crazy to want to use in > domain names. I think you meant hostname here. Remember that hostname rules apply to delegations. Domain name rules are a broader category which also include things like SRV, email addresses, and so forth. > I asked long ago, shouldn't we try to design a solution that handles > both domain labels and email address local-parts? I was told no, > that's out of scope, a different working group will take care of > email address local parts. I share your concern, but I got the > impression that this was not open for debate. The email people will come up with their own character restrictions, which is exactly the point. The namespace should ALLOW them to come up with mapping rules, case-folding, normalization and whatever other rules they wish. I mean, if they want to have a perverse mailbox sequence, why shouldn't they? The codecs should only deal with inputs and outputs. I'm pretty certain that's what you were told. -- Eric A. Hall http://www.ehsco.com/ Internet Core Protocols http://www.oreilly.com/catalog/coreprot/
